The goal of this proposal is to develop protease-activatable vectors that can be used to target the delivery of therapeutic genes to sites of active angiogenesis. We have developed a novel strategy for targeting retroviral vectors through protease- substrate interactions. We also made a library of replication competent substrate retroviruses presenting a diverse array of potential protease substrates whose cleavage would result in their activation. By propagating the library on selector target cells we isolated viruses that were activatable by endogenous cellular proteases. We now hypothesize that the spectrum of protease activities known to be associated with tumor angiogenesis can be exploited by making protease-activatable vectors that target gene delivery to sites of active angiogenesis. We propose to use an in vivo library selection strategy to obtain viruses that are activated by the proteases of angiogenesis. We will first build a new library of replication competent retroviruses whose infectivity will be strictly dependent upon cleavage of a diversified hexapeptide linker sequence. We will then select viruses from the library on the basis of their ability to propagate in the stroma of human tumors xenografted in mice, and to spread via the bloodstream from one tumor xenograft to another at a distant location. The protease susceptibilities of the envelope glycoproteins of the selected viruses will then be characterized and they will be used to construct nonreplicating retroviral and lentiviral vectors that target gene delivery to proliferating tumor blood vessels in vivo.